Kong has expanded its Kong Konnect platform to provide governance, observability, and cost management for autonomous enterprise AI agents, addressing the growing complexity of multi-step digital workflows. The update introduces new capabilities to secure identity, track spending, and manage network traffic as corporate operations shift from simple model calls to complex agentic interactions.
Managing the Cost of Agentic Workflows
Modern agentic deployments operate across complex paths involving multiple models, tools, and peer agents, moving beyond traditional generative AI architectures that concentrated on sending individual prompts directly to large language models. Dynamic agentic workflows interface directly with enterprise data repositories, external tools, and software applications to execute tasks on behalf of users.
Augusto Marietti, Chief Executive Officer at Kong, stated that AI is rapidly introducing connectivity problems for enterprises. He noted that an agent creates value when it can securely work with APIs, tools, data, applications, and events. However, every interaction must be discovered, authorized, governed, observed, and paid for.
"Kong AI Management brings those concerns together into one connectivity layer, so enterprises can give AI access to what it needs while controlling how it is used," Marietti said.
Shafii emphasized that managing these connections securely, reliably, and efficiently is essential to turning AI investment into innovation. He described "AI Connectivity" as a foundational part of infrastructure, where agents connect to models, data, and actions through APIs and protocols such as the Model Context Protocol (MCP). Events provide real-time information about the world, allowing systems to react dynamically. Without structured operational oversight across these integrations, expensive compute capacity, inference resources, and engineering hours are squandered.
New Security and Cost Controls
The expansion includes Kong AI Gateway 2.3, which broadens runtime management across digital networks by enforcing corporate policies and supervising execution costs across model and MCP traffic. This runtime control allows engineering teams to maintain operational consistency across evolving multi-model environments.
To strengthen security perimeters, the platform introduces Kong Identity, which establishes continuous identity tracking across staff members, programmatic applications, and autonomous agents. This unified verification model helps organizations understand precisely which entity is acting across enterprise traffic to apply strict authorization rules.
Kong Token Vault protects corporate networks by brokering downstream credentials at runtime rather than requiring agents to store keys directly. This runtime credential management curbs token proliferation across AI applications while supporting identity-based governance rules.
Complementing these security mechanisms are cost management capabilities that connect infrastructure usage to specific departments, business activities, and autonomous agents. Enterprise teams can establish formal budgets, operate showback or chargeback accounting, and deploy optimization techniques such as caching, prompt compression, and model routing.
Advanced AI Observability delivers end-to-end visibility across complete operational sessions, tracing interactions across models, tools, agents, and protocols. Technical teams can audit individual conversational turns to investigate latency spikes, operational errors, and cost anomalies.
Centralized Access and Partner Ecosystem
Kong also introduced AI Registry to provide a governed index of available digital capabilities across the business. This catalogue enables organizations to define the precise ecosystem that autonomous systems can discover, establishing access parameters for business models while directing tool usage during automated tasks. By maintaining centralized control over every software endpoint, the system ensures that dynamic agent workflows only interact with permitted enterprise resources.
Kong collaborates with several enterprise technology partners across its connectivity and API ecosystem:
- Amazon Web Services (AWS): Supplies on-demand cloud infrastructure, compute capacity, and machine learning services for scalable enterprise deployments.
- Microsoft Azure: Provides enterprise cloud environments, AI compute clusters, and hosted model tools for corporate software applications.
- Google Cloud: Delivers hyper-scale cloud hosting, infrastructure resources, and enterprise AI capabilities to power distributed software services.
- Persistent Systems: Offers global digital engineering, enterprise system integration, and software modernization services across corporate IT networks.
- NTT DATA: Delivers multinational IT consulting, systems architecture, and managed cloud services for large-scale enterprise transformation.
- ServiceNow: Builds digital workflow platforms that automate enterprise business processes, IT service delivery, and corporate operations.
The company’s growth trajectory reflects its increasing importance in the infrastructure space. Founded in 2009 as Mashape in Milan by schoolmates Augusto Marietti and Marco Palladino to develop no-code application tools, Kong open-sourced an internal API gateway in 2015 to decouple its core marketplace into microservices. After selling Mashape in 2017 to focus entirely on developing Kong, the company expanded its staff beyond 100 members.
In 2021, Kong launched the Kong Konnect platform and achieved unicorn status following a $100 million Series D round valued at $1.4 billion. By 2023, the company had exceeded $100 million in annual recurring revenue and grown its headcount past 500 staff members. The 2026 expansion into agentic governance marks the latest phase in securing and observing autonomous enterprise interactions.
Also from the source material: Kong’s ecosystem includes Confluent, which provides Apache Kafka-based data streaming capabilities for real-time event processing within these enterprise networks.
Comments
No comments yet. Be the first.